AUSTIN, Texas — A group of Texas Republicans wants to audit the 2020 election results in just the large, mostly Democratic counties across the state. If they get their way, they'll miss many of the real — but minor — errors in the state's vote count.
That's according to a team of researchers that conducted a statewide analysis of the results across both Democratic and Republican counties. The group found a series of errors that would not come close to changing Republican Donald Trump's victory in the state or any other statewide race. But the errors stretch across both Republican and Democratic counties.
The research adds to a pile of evidence that contradicts the belief, widespread among Republicans, that elections in Democratic areas are rife with errors, irregularities and mismanagement. While errors in the tally do occur, research shows they tend to be random and small scale and do not benefit one party or the other.
In Texas, the mistakes, detected by election researchers from the University of Florida, were scattered across 37 of Texas' 254 counties. They added or subtracted a handful of votes from various candidates with no skew toward one party or the other. Trump apparently received 223 more votes than the 5,890,347 that the Texas secretary of state lists as the Republican's total. Democrat Joe Biden appears to have received 155 more votes than his listed 5,259,126, according to the research.
Minor mistakes like the Texas ones are relatively common, say election experts. In Texas, the errors are likely due to the state's use of an older computer system that requires counties to enter their tallies by hand, increasing the risk of errors when the wrong digit is typed.
But they take on greater significance at a time when Trump supporters are calling for increased scrutiny of Democratic election offices, such as in Texas, where Attorney General Ken Paxton is among the Republicans supporting legislation that would audit the state's largest counties — most of which were carried by Biden last year.
"If Texas is going to focus on the blue counties, that's probably the wrong thing to do," said Michael McDonald, the political scientist who led the team that found the Texas discrepancies. He also operates the U.S. Elections Project, which has tallies of all national elections since the nation's founding. "They should look at all the counties because there's something broken with this system."
Texas' elections director, Keith Ingram, said the state is working to upgrade its computer system to one that will automatically transmit counties' final tally to the state, eliminating the need for local election directors to type it in. But that may not be ready for testing until next year's primary.
"It's not a huge number," Ingram said of the discrepancies, which he had not heard of until contacted by The Associated Press. "Obviously, we want everything to be precisely spot on, and we're going to give counties the ability to do that."
Election experts say the Texas example shows the need for regular oversight of the system — as opposed to the so-called forensic audits that Trump supporters have backed that don't follow established procedures and chase wild theories of voting fraud. An audit authorized by the Republican-controlled state Senate in Arizona's largest county, for example, conducted a search for bamboo fibers on the ballots after a conspiracy theory circulated that fraudulent ballots for Biden were shipped in from Asia.
The Arizona audit has become a rallying cry for Trump backers who are convinced the election was stolen, although repeated professional audits and recounts have shown no mass voter fraud. They are pushing for new audits in Pennsylvania and Michigan. In Wisconsin, the Republican head of the assembly's elections committee has promised an audit. In Texas, nearly three dozen Republicans signed onto a bill that would audit 13 counties with populations of more than 415,000 people. The bill has been in limbo, along with other GOP priorities, since Democrats bolted to Washington to block new voting laws.
The Republican lawmaker who proposed the Texas audit, state Rep. Steve Toth, expressed bafflement to The Washington Post about expanding it to smaller counties. "What's the point?" Toth said. "I mean, all the small counties are red."
Toth's office did not respond to emails seeking comment from The Associated Press.
The Texas discrepancies are "why having real election audits and reconciliation audits is important — to make sure any transcription error like that is caught," said Tammy Patrick, a former Maricopa County election official who works with The Democracy Fund and is a critic of the partisan audits. "It's important that we don't look at those kind of errors as cavalier or that they don't matter, because they do, but they're not indicative necessarily of fraud."
The mistakes range from Travis County, which includes Austin, reporting 100 too-few votes for a state Supreme Court candidate to small, conservative counties like Gillespie in Texas' Hill Country, where the totals were a single vote off in a handful of races.
"If it's different, then that's going to be an input error, and it's probably on the secretary of state," said Terry Hamilton, a clerk in the Gillespie County elections office.
The discrepancies don't all stem from transcription errors. In Kleberg County, near the southern Texas coast — a 30,000-population county won by Trump — Election Director Stephanie Garza had long-delayed knee replacement surgery three days after the election.
That meant she wasn't able to personally ensure that ballots judged to be legal days after the election — known as provisional ballots and, in Texas, also limited ballots — were included in the final numbers sent to the secretary of state.
Because they were omitted, Trump's winning total was understated by 71 votes, and Biden's total was missing 67.
Garza said she was mortified to find out about the omission, even though it represents only a minuscule fraction of ballots cast in the presidential race.
"As an election director, you don't want to be off whatsoever," Garza said.